Half a Million Girls in Wells, and We Still Change the Channel

Lex Fridman opens with a blunt question: what is more effective at making a net positive impact on the world, empathy or reason? Harris says it depends on which empathy you mean. Cognitive empathy, as he describes it, is almost a species of reason — an accurate theory of mind that lets you understand why someone is suffering or happy and navigate the relationship. A second layer, not incompatible with the first, is what people often mean by the word: emotional contagion. You feel depressed because the person next to you is depressed. Their problems bleed through.

Both capacities matter, he says, but the contagion piece is often a bad guide for ethical behavior. The thesis is not originally his. He credits the psychologist Paul Bloom and the book *Against Empathy*. The clearest exhibit is the identifiable victim. A little girl falls down a well; you see the parents on television; you hear her cries. Harris recalls wall-to-wall CNN coverage, maybe 72 hours, of extracting a child from a well some 20 or 25 years earlier. The country stops. Money and compassion pour in.

Tell people instead that a genocide is raging in a country they have never visited and never intended to visit, and the numbers do not make a dent. The story is boring. They change the channel. Perversely, he says, it could be 500,000 little girls fallen down wells in that country and we still would not care. Many of us have come to see this as a bug in moral psychology rather than a feature. For large decisions about suffering and what to protect, Harris thinks reason is the better tool — without wanting to throw empathy away.

Fridman presses the uncomfortable implication. Harris has talked about effective altruism; would he rather live in a world where people care more about their own children than about a hundred strangers? Harris is not sure there is one right peak on that part of the moral landscape. The opposing ideal, as articulated by the Dalai Lama or classical Buddhism, is true dispassion toward friends and strangers. The mind of the Buddha would love all people equally. By that light, spending nearly all of one's energy on one's own kids while a genocide in Sudan takes less than 1% of bandwidth looks like a failure to actualize compassion.

He is not sure it would be a better world if everyone ran "the Dalai Lama program." Some ethical prioritization of one's nearest and dearest might be optimal, he argues, because we will all be doing that inside norms, laws, and institutions that force dispassion where it matters. When his daughter is sick, he wants her to get attention, and he worries about her more than about everyone else in the lobby. He still does not want a totally corrupt hospital — one that treats her better because she is his daughter, because he bribed the person at the door, or because that person is a fan of his podcast.

Press people down Maslow's hierarchy, individually and as a society, and those variables change for the worse. In a collective emergency you get a lifeboat problem: everyone is scrambling, and fairness, norms, and the vestiges of civilization get stripped off. Harris's warning is that we cannot reason from those emergencies back to ordinary life. In ordinary life we want justice and fairness, not because we are saints, but because in game-theoretic terms we tend to be better off in a fair system than a corrupt one — even while the spotlight of concern stays on family and friends.

Fridman asks what empathy is worth when half the country disagrees about immigration, the pandemic, or whether an election was fairly run. Harris says it is important and only takes you so far. "It doesn't get you to truth." Truth is not settled by democratic principles. People believe things for understandable reasons that are nonetheless bad reasons — they do not scale, they do not generalize, and they are not reasons anyone should adopt epistemologically. You can still care about the circumstance that produced the belief.

His running example is Donald Trump. Harris has been harshly critical of Trump and of people who endorsed him, or who stayed silent, once a sitting president would not agree to a peaceful transfer of power. He can still understand why millions of frustrated people, many of them victims of globalism and other trends, bought the myth of the brilliant businessman. They had watched him on television for nearly a decade and a half pretending to be a genius who could get things done. Harris says he does not necessarily blame those millions. He still bemoans the phenomenon as a very bad state of affairs. The better word for what he hopes to embody, he adds, is compassion: really wishing strangers well, and noticing that there is no final opposition between selfishness and selflessness. You do not actually want a society where you have everything and most other people have nothing.

Fridman steel-mans the other side: perhaps Harris bought the myth of a government of the people, while Trump supporters see a corrupt elite bureaucracy and a chaos agent who, however flawed, might shake it. Harris agrees with the first part. He does not pretend the United States is an idealized representative democracy, and he is not even sure everyone's opinion should have equal weight on every topic, including his own. He declines some conversations because he thinks he is the wrong person to have them. He still calls electing Trump a very bad bet: decades in public view of gross incompetence and a malevolent selfishness, a man missing basic moral and psychological tools. The analogy he likes is an evil Chauncey Gardiner, the empty gardener in *Being There* whom Washington mistakes for a sage. Trump's actual genius, Harris says, is creating a spectacle around himself — "the pure pornography of attention," which only works if you are shameless.

Fridman asks Harris to steel-man the meme that Trump broke his brain — Trump derangement syndrome, the claim that the reaction was out of proportion. Harris calls it a clever meme because it throws the problem back on the critic. In truth, he says, the true syndrome was not seeing how dangerous it would be to promote someone like Trump, and then, in the final moment, still supporting a sitting president who would not commit to a peaceful transfer of power. If that was not a bright line, "you have been deranged by something." He calls it one minute to midnight for American democracy.

That the country did not fall into a full constitutional emergency after January 6 is not, for him, evidence that the worriers exaggerated. He says you can count on two hands the people who held things together, and that if Mike Pence had refused to certify the election the crisis would have been real. Trump did not succeed for lack of trying. Harris's analogy is nearly being run over by a car: the fact that you were not hit does not mean the surge of adrenaline was a mistake. He is not spending much time yet on a possible 2024 return; he is waiting until the moment requires thought.

He understands the other posture: Trump as funny, entertaining, a stick in the wheel of business as usual, nothing much at stake if NATO is slighted or if he claims to trust Putin more than American intelligence. Harris can inhabit that space for a few minutes. The deeper worry is that the culture is entertaining itself to death — goofing around at scale on problems where lives hang in the balance. Institutions have lost trust, sometimes for good reason, including capture by ideologies that are not truth-tracking. He still thinks the answer is to rebuild them and to value expertise over amateur speculation, conspiracy thinking, and bullshit.

Fridman offers another frame: perhaps Trump is not an ultra-competent malevolent executive in the line of Lenin or Hitler, but a jokester whose chaos revealed how fragile the system is. Harris largely accepts that. He does not think Trump is a Putin-like or Hitler-like figure. Trump is not ideological and does not care about anything beyond himself, which makes him much less scary than a genuine totalitarian. Asked whether that is more *Brave New World* than *1984*, Harris says this is the analogy Eric Weinstein never stops badgering him about — and that Eric is still wrong.

The truly depressing and scary thing, he says, was a country in which at least half the public saw nothing wrong with elevating someone who obviously did not know what a president should know and who could not even pretend to care about other people in a credible way. If there is a silver lining, it is the one Fridman sketched: the episode showed how vulnerable the system would be to a truly brilliant, conniving, ideological figure. Trump did not exploit those holes. Someone more competent might. Harris says he did not previously understand how much the system relies on norms rather than laws. Civility, almost. It is quite possible, he adds, that Trump never did anything illegal in the sense that should put him in jail — that remains to be seen. The chaos, the drop in stature, the years spent on nonsense, were largely norm violations. It is not illegal for a sitting president to refuse to commit to a peaceful transfer of power.

Fridman asks whether those violations were useful perturbations, a way to learn the limits of civility. Harris says this was not humorous. Trump did everything he could to steal the election, declaring it fraudulent in advance, trying to stop counts he knew would be disproportionately Democratic because of mail-in ballots, and sending Rudy Giuliani and Sidney Powell into a circus that included the Four Seasons landscaping episode. Republican election officials and judges threw the claims out. Trump was not competent, and neither were the people around him; that is why it failed, not because he did not try. People with Gulfstreams who could fly to New Zealand if it all went to hell were, in Harris's view, in no position to cheerlead the nihilism.

He will grant Trump is funny, a good entertainer, and that he agreed with much of the policy impulse — taking China seriously as an adversary, controlling the border — perhaps 80% of the concerns Trump's fans celebrated. That is not the issue. The issue is a superpower of deranging almost everything he touches and compromising the integrity of almost anyone who comes into his orbit. Harris wants a house cleaning of government. He is not a fan of the propagandized phrase "deep state," but he thinks something like it is too rigid. He wants that job done by smart, well-informed, well-intentioned people, not "literally the one most selfish person" anyone could name.

That is the context for Harris's infamous line about Hunter Biden's laptop. He says he can claim, with confidence, that he does not care what was on it — including any evidence of corruption by Joe Biden — because precious little of that has actually emerged, and because even obvious corruption would not balance the scale of what is already known about the two men. Both have lived in public for almost as long as Harris has been alive. Show him that Biden has a secret life of Bugattis and Andrew Tate decadence, and he will start to revise. On the evidence in view, he says, there is no comparison. In 2020 it was a forced choice between a sitting president who would not commit to a peaceful transfer and a man obviously too old, with a crack-addicted son who lost a laptop. He was going to take Biden "in spite of whatever litany of horrors was gonna come tumbling out of that laptop." Fridman quotes the dark joke: Hunter Biden could have had the corpses of children in the basement. Harris's ratio is a firefly to the sun.

Fridman objects that this is surface scandal. The military-industrial complex can spend hundreds of billions, even trillions, on wars in the Middle East, kill civilians, and generate hatred, all while everyone is polite and there are no strippers or cocaine on a laptop. Harris does not deny the underlying point. His claim about Trump is not that strippers are the worst crime. It is that a person calibrated only to wealth, power, and fame will not track environmental risk, pandemic risk, or nuclear proliferation. He notes that Trump repeatedly asked generals why the United States could not use nuclear weapons, and why he had fewer of them than John F. Kennedy, "as though that were a sign of anything other than progress."

On the military-industrial complex itself, Harris says Fridman is right, and that the phrase comes from Eisenhower. So much of what ails public life is bad incentives, which corrupt even good people and corrupt bad people more. The goal is to get reasonably non-pathological people into the system, and better still to invert the incentives so that even bad people behave as if they were good. He was never a supporter of the Iraq war; he was a supporter of Afghanistan, which in hindsight looks at best ambiguous and more likely a fool's errand. Some of the neoconservative faith in nation-building at gunpoint has changed for him. Girls going to school after the Taliban is the right side of history; trillions of dollars could not solve a problem of religious maniacs willing to die, and let their children die, for seventh-century norms.

Fridman asks whether the Twitter Files, and the Hunter Biden laptop story, bother Harris — a collection of people deciding whom to ban, open to bias. Harris thinks that bias will always be there in the absence of perfect AI, and that Twitter is not as important as people think. Leaving the account was, in his experience, unambiguously good. Even the useful parts — a newsfeed of about 200 smart people, articles to boost — fragmented his attention. Almost 100% of the professional controversies that bent his life over the previous decade, he says, were Twitter. Blowback to a podcast would launch a cycle of response that he now sees as mostly specious. The sane way to use the platform is as a one-way marketing channel. He wanted dialogue, kept coming back for it, and noticed it was changing how he felt about people he knew and people he did not.

He describes a funhouse mirror that turned ordinary people into sociopaths, the social-media version of road rage. Friends he could sit down to dinner with were coming at him in ways they never would in person. He deleted the account because it was making him a worse person. He is agnostic about whether Elon Musk, who had just taken over, or anyone else can make a healthy social platform; he thinks Twitter is obviously negative for Musk too. The Hunter Biden clip that exploded was taken from a podcast and, Harris says, highly misleading even though the full context was not his finest speaking. Most people see a clip and assume no relevant context is missing. He is careful to add that he gives Trump the same benefit of the doubt: the famous Charlottesville "good people on both sides" clip, promoted from Biden on down through the *New York Times* and CNN, did not capture what Trump said about the Nazis. Harris says he has always been clear about that, and that he has not heard an honest apology from those who used the clip.

The next step, he warns, is manufacturing clips with AI and deepfakes, where most people most of the time will not even know whether they are looking at something real. That is a misinformation apocalypse on the cusp. Fridman wonders whether humor, Trump's absurdism, is the only remaining defense. Harris says we cannot live that way. People act on what they assume is true — which medicine to give a child, whether a comet is coming. He expects fake journal articles, already being drafted with ChatGPT, that within a year would take a PhD to debunk. The people celebrating that cataclysm, he thinks, either have nothing to lose or have convinced themselves they are insulated.

How we relate to authority in science, Harris says, looks like a paradox and is not one. The truth of a statement does not depend even slightly on the credentials of the person making it. A Nobel laureate can speak nonsense; a backyard astronomer can find a comet the professionals missed. It is also true that most of the time real experts are much more reliable than frauds, and that expertise is a thing. You do not want randos in the cockpit of a plane in a storm. Public health, Ukraine, climate change — we are at 30,000 feet a lot of the time, the clock is ticking, and the solutions are non-obvious. Deferring to experts, at minimum, prevents spectacular errors of incompetence.

When times are good, he views "do your own research" as an invitation to waste time. Times are not that good. Experts are corrupted by bad incentives, institutions have been captured by political ideology that is not truth-tracking, and the woke encroachment into universities, journals, and government has been, in his word, deranging. None of that erases the gulf between someone who has done the work and someone who has not. Harris has a PhD in neuroscience and can read a medical paper; if someone close to him gets cancer, he still will not pretend to be an oncologist. Get a second and third opinion, he says — from real experts, not from Robert F. Kennedy Jr. There is no scenario in which you should take your view of vaccine safety, climate, or Ukraine from Candace Owens. The doctor who drank *H. pylori* and overturned the stress theory of ulcers shows that consensus can be wrong. It does not show that expertise is a fiction, especially in an emergency.

On COVID, much of Harris's concern was never narrowly about how dangerous the illness was. After the first months, it became a worry about a failed dress rehearsal for a truly scary pathogen. Society had built informational tools it did not know how to use and enrolled itself in a deranging psychological experiment that makes coordination almost impossible. He thinks there was a period when, given the data and the expectation that the vaccines would knock down transmission, it was rational for basically everyone to get vaccinated, and a civic duty because some people cannot. He waited in line at 4:00 a.m. for overflow doses. He also says the target moved: safety concerns around mRNA shots for young men, widespread natural immunity, a less dangerous virus. He does not know that he will get the next booster. At every point, he insists, he was the wrong person — and so was Bret Weinstein — to have strong opinions. You can always find a PhD or an MD to champion a crackpot idea; people once sat before Congress and denied that smoking caused lung cancer.

He still thinks it was rational, early on, to attempt some lockdown to bend the curve and spare hospitals, given Italy. In retrospect, a society this allergic to the demand may simply be unable to lock down unless the pathogen is much more deadly. A malicious clip of him on that point was made to sound as if he wished more children had died; even the clip, he says, betrayed the context. The person cutting those clips was someone Weinstein signal-boosted. Harris's regret is larger than COVID: he has not sorted out the ethics of criticizing friends and acquaintances who have public platforms and are, in his view, behaving badly. With Fridman's Kanye interview he reached out privately first, then said the same thing on the air. With Elon Musk, at least at the beginning, he tried the same. With Weinstein, he thinks he failed to make enough private effort before it spilled into public.

Fridman defends what Harris had criticized as a naive philosophy of love: sitting with Kanye West, getting through the minefield, maybe hugging at the end, and showing the world a willingness to listen. Harris says the cases are not analogous. He has never met Kanye, but he thinks the man is in a mental-health crisis, a colossal asshole, or both. If it is mental illness, putting a microphone in front of him and letting him go off about Jews is not a favor. If it is ordinary antisemitism, then the job — which Harris says can still be done with love — is to make it clear, to him and to the audience, that he is full of shit. Compassion does not always have a patient, embracing face. There is a sword of wisdom. Harris would not have aired the conversation.

Leave mental illness aside, and Harris finds the most offensive note not even the antisemitism but the delusional self-estimate: an omnibus genius to rival or exceed Shakespeare who then talks for hours without saying anything interesting, insightful, or factually illuminating. It is, he says, a Trumpian sort of delusion, except that Kanye seems to believe his own press. Coupled to antisemitic memes, Harris thinks that raises rather than lowers the odds of violence — he mentions the "knockout game" against Orthodox Jews in Brooklyn, and an ambient antisemitism in parts of the Black community. He is Jewish, not religious, a student of the Holocaust, and says that in his own lifetime he has almost never taken antisemitism very seriously. He thinks that is a mistake. It is a unique hatred, knitted to conspiracy theories, able to animate the far right (Jews as not white) and the left (Jews as extra white, a model minority used against other minorities).

On platforming more generally, Harris describes an uncanny valley. If someone is sufficiently evil — if you are interviewing Hitler — you need not signal disagreement; you can go in as an anthropologist of evil, and he thinks you always do that interview once the man is legitimately Hitler. On the cusp of power, the question is whether you can take the wind out of their sails without giving them oxygen. Some people pack so many confabulations into every sentence that you cannot clean the mess in real time. Alex Jones is his exhibit: a string of lies and micro-truths at a velocity no conversation can falsify. Harris spent years privately trying to deprogram a family member on 9/11 conspiracies and learned the structure: not a thesis but a proliferation of incompatible anomalies, perfect competence attributed to thousands of secret murderers, and a Whac-A-Mole you cannot win. That is the pattern he feared in the COVID fights, and why he would not interview Trump. He thinks Joe Rogan was wise not to have Trump on. One big laugh in a debate, Harris says, and you win. We are monkeys, and we like to laugh.

Fridman notes that Rogan, on a podcast with Jordan Peterson, said they still have hope for Harris. Harris had already reached out privately. He considers Rogan a friend, an extraordinarily ethical person, and has no policy against doing the show. The lanes are different. Rogan is a standup comic who can pull the rip cord — I'm a comic, we were stoned, don't take medical advice from me. Harris cannot. The larger mechanism he worries about is audience capture, a phrase he credits to Eric Weinstein. When part of Harris's audience hated his Trump criticism, he leaned into it; when another part thought he exaggerated wokeness, he leaned into that too. He did not want those cohorts. He thinks something like capture is the only way to explain how Weinstein, with the whole universe to talk about, did a hundred podcasts in a row on COVID.

The cartoon version of his objection, which he says he voiced about 18 months earlier and still holds: even if Weinstein turned out to be right, even if ivermectin were a panacea and mRNA vaccines killed millions of people, he was still wrong at the time. The reasoning was flawed, the facts were bad, the confidence was unjustified. That will always have been true of that moment. Harris would be powerfully annoyed if teenage boys were being forced to take mRNA shots given myocarditis concerns; he also cites a figure that the United States lost something like 300,000 people unnecessarily to vaccine hesitancy. He platformed people such as Eric Topol and Nicholas Christakis. He says he did about two COVID podcasts, not a hundred. Once it was clear how to live with the virus, he lived with it. The thing he did not do was make it his life. What still occupies him is next time, when something orders of magnitude more dangerous hits a public that no longer trusts a public-health regime.

Asked what he admires in Weinstein, Harris does not hesitate: very smart, very ethical, wants good things for the world. Being crosswise on COVID does not make him a bad person. The worry was messaging at scale and a responsibility not to get people killed. On lab leak, Harris was never skeptical; his early position was that it did not matter next to what to do about the pandemic. The virologist Alina Chan changed his mind on timing: the origin is best investigated immediately, before evidence disappears. He still thinks either story makes China look bad, and that a wet-market origin is in some ways more politically invidious — barbarian, another century — whereas a lab leak could happen to any advanced country running dangerous labs. Fauci's exchange with Rand Paul looked lawyered and shady. Harris thinks gain-of-function research should probably be banned, and he no longer even understands virus hunting given how fast genomes can be sequenced and vaccines designed. Being early on lab leak, he insists, does not make someone right about ivermectin or mRNA, and it is possible to be right by accident.

Harris likes Elon Musk, had a lot of fun with him, and is a fan of almost all of Tesla and SpaceX. He thinks that if many of Musk's aspirations are realized the world would be better. What he objects to is the Trumpian register of Musk's own behavior on Twitter: forwarding a conspiracy that Nancy Pelosi's husband had been attacked in a gay tryst, from a site that had claimed Hillary Clinton was dead and a body double was campaigning; suggesting a former employee, Yoel Roth, was a pedophile, after which Roth was inundated with death threats. Scale matters, Harris says. Doing this in front of 130 million people is not a joke. Musk brought Kanye on knowing the antisemitism and mental-health history, then kicked him off for a swastika — a ban Harris is not even sure a free-speech absolutist can enforce.

The break over COVID came very early. In March 2020 Musk tweeted something like "panic over COVID is dumb," and it became the most engaged tweet of the week, bigger than anything Trump was posting. Harris thought people were reading genius-technologist inside information into a shrug, and that it was landing as "non-issue, it will peter out." He reached out privately, from a place of love, trying to get Musk to walk it back. The intervention did not work. At that point, Harris says, tens of thousands of American deaths were already a floor, and Nicholas Christakis had predicted on Harris's podcast on the order of a million. Where things broke down was not over masks or lingering uncertainty. It was over whether there was a water bottle on the table.

Musk was formative in Harris taking AI seriously; they went together to the early Puerto Rico conference. Harris's concern about AGI is unchanged since his 2016 TED talk: unless a self-improving superintelligence can be permanently tethered to human values, we become ants. He does not believe anyone has figured out how to do that, or whether it is possible. His guide is Stuart Russell and the book *Human Compatible*: do not specify the value function up front; keep the system uncertain about what humans value and always trying to approximate it. Harris sees problems even at a high level. Which humans get a vote? What about revealed preferences — Twitter, the nuclear brink, hot girls in yoga pants on TikTok? Russell's observation about YouTube is that the algorithm does not only cater to preferences; it changes people to make them more predictable. The worry is not cartoon malice. It is instrumental goals, the way we pave a driveway without forming an intention to harm insects.

Imagine creating a race of people ten times smarter, he says. They begin to talk about things we do not understand, want things we do not understand, and view us as obstacles. We become the chickens. He buys the science-fiction claim that this is existential if we do not do it well. Narrow systems such as ChatGPT already threaten the information space before they have goals of their own. He has no timeline intuition except two assumptions: substrate independence, and that we keep making progress. From people who know more than he does, he has also gathered reasons to doubt that current deep learning is sufficient for AGI — the systems do not generalize like children, and they have strange holes. Fridman notes that the holes keep shrinking.

Eric Weinstein had told Rogan that he and Harris were contacted about UFOs. Harris says they were contacted by the same person, who seemed credible, promised Zoom calls with people they would recognize — former CIA heads — and never cashed the check. Harris lost interest. He ran a thought experiment: if alien craft were real, it would be the biggest story in anyone's lifetime, and yet it barely rose into his attention next to COVID and Trump. He treats that partly as a defense against a history of bogus claims. He mentions Mick West's YouTube debunkings of some of the videos that a credulous *60 Minutes* piece had aired. Extraordinary claims require extraordinary evidence. The AI conversation is different because we would be designing the aliens, and he is not skeptical that we will eventually succeed.

The girl-in-the-well problem returns as a civilizational one. We cannot make certain information emotionally salient. A 30% chance of nuclear war is Russian roulette with a three-chambered gun aimed at everyone's children, 24 hours a day, and it still loses to a vivid local story. Harris cites people such as Graham Allison and William Perry putting the risk of a proper nuclear war in the next generation around 50%. The status quo is old ICBMs aimed at the biggest cities, vulnerable to a lunatic or bad data; he says we have been saved at least twice by Soviet officers who would not pass a false alarm up the chain. Deepfakes of a president claiming a first strike make that worse. The response may be more siloing: do not trust a photograph unless something like Getty Images can attest it, if that is even technically possible.

Asked by Ron Howard, via Twitter, about society beyond Mars, Harris's picture is humbling. There will be a first fistfight on Mars, a first murder, extramarital affairs. Only the spacesuits will limit how badly apes can behave. He expects a pendulum swing back toward face-to-face life; after leaving Twitter he noticed friends who had never been on it and had never understood what he was doing there. All we have is attention. The ground truth of wellbeing, he says, is the meditative recognition that consciousness in the present is already a circumstance of freedom — you can be happy now, before the next desire is gratified. Free will is one of the thoughts that commandeer that freedom. When he is mindful, there is no one in the middle of the boat; there is just the river. Voluntary and involuntary still differ. The intention to reach for a glass still just arises.

He has been kicking around a further claim: maybe only the actual is possible. Every regret and every worry depends on a spooky surplus of paths that did not happen. Worry, he thinks, is almost never psychologically appropriate: if you can do something, do it; if you cannot, you are only suffering twice. He wants to talk to a physicist about how this sits with many-worlds and with Einstein's block universe, in which the last page of the novel exists as much as the first. That sounds confining and is, as he lives it, a circumstance of pure discovery. Hope, at the end, comes back to character and incentives. Most people are good and converge on the same core values; Twitter had convinced him otherwise, which is why he left. There is enough wealth on this one rock. The asymmetry he fears is that it is easier to break things than to fix them, and that technology now lets a minority derange millions — gain-of-function know-how, nucleotides in the mail. His unpopular free-speech stance is that if you own Twitter you should be able to kick off the Nazi you do not want to platform; that is the owner's speech. He thinks we are in a bottleneck. If we get it right, there is no limit to how happily billions of creative people could collaborate.